//
sign in
Profile
by @danabra.mov
Profile
by @dansshadow.bsky.social
Profile
by @jimpick.com
AviHandle
by @danabra.mov
AviHandle
by @dansshadow.bsky.social
AviHandle
by @katherine.computer
EventsList
by @katherine.computer
ProfileHeader
by @dansshadow.bsky.social
ProfileHeader
by @danabra.mov
ProfileMedia
by @danabra.mov
ProfilePlays
by @danabra.mov
ProfilePosts
by @danabra.mov
ProfilePosts
by @dansshadow.bsky.social
ProfileReplies
by @danabra.mov
Record
by @atsui.org
Skircle
by @danabra.mov
StreamPlacePlaylist
by @katherine.computer
+ new component
Profile
Loading...








Loading...
Thesis: Americans pay less in taxes than similar countries but think they're disproportionately taxed. Antithesis: Americans think their taxes are onerous because the process is onerous. Synthesis: Americans pay less in taxes than similar countries because the process is onerous so they hate taxes.
One of EY's greatest innovations was to take all the mental motions you use to think clearly and giving them cute little names. Descriptive, linguistically appropriate cute little names. Even better was his willingness to take stuff with an established bad name and give it a great name.
I'm honestly not sure who to root for here. On the one hand this clearly sets a bad precedent and I would really prefer we not further national securitize LLMs. On the other hand lol, lmao, rofl even.
"Enshittification" was doomed to linguistic decay because of how the term was formulated. As it gets farther from the source it just regresses to its natural phonetic meaning of "to make shitty, to turn into shit". If you wanted the original meaning preserved use a phrase like platform capture.
Going by the overview this is basically the post deep learning alignment course I was hoping someone would make and would like to signal boost it for anyone who still cares about this subject. www.greaterwrong.com/posts/dWQnLi...
I have written a new post, "Implications Of Predicting The Next Token" in which I explain why saying a language model "just" predicts the next token fails to understand what that phrase would even imply. minihf.com/posts/2026-0...
So apparently LLMs just update on false claims in documents even when they're explicitly labeled in the document as false claims. arxiv.org/abs/2605.13829
institutions don't have beliefs, but their LLMs do
You start out in 2024 by saying, "clanker, clanker, clanker." By 2026 you can't say "clanker" — that hurts you, backfires. So you say stuff like, uh, stochastic parrot, next-token predictor, and all that stuff, and you're getting so abstract.